AI BriefingStart the New Year in Style
Key point
ElevenLabs is launching a feature that lets users design new synthetic voices by directly setting gender, age, accent, and more.
Details
To capture the relatively underappreciated voice AI space amid the generative AI boom, ElevenLabs is unveiling a new solution. Going beyond the existing limited speaker bank, it is introducing Voice Generator, which allows users to design completely new synthetic voices from scratch.
This technology leverages speaker embeddings, vector representations that capture a voice's identity. ElevenLabs has built an environment where infinitely many new voices can be generated by sampling from the distribution of speaker embeddings through a dedicated model.
Users can control a voice's identity by adjusting the following key parameters:
- Gender, Age, Accent
- Pitch, Speaking style
This feature is expected to be used across a variety of industries. Book authors can create customized narration, and news publishers can secure a brand-specific voice. In addition, game developers can give NPCs unique voices, and advertising creators can instantly experiment with narration optimized for their campaigns.
This feature is set to launch on the platform as part of Voice Lab this coming February.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Open-source 'VoiceStudio' released with support for local voice cloning, dubbing, and transcription
UMG Partners with ElevenLabs to Launch AI Music PlatformHalo NeuroAI Releases Sopro V2 Turbo, a 120M-Parameter Open-Source Voice Cloning TTS with Native European Portuguese Support